LLMs used to hallucinate a lot. It was fairly common. However, when working with supporting documents that the LLMs can reference as context, I haven't seen an LLM hallucinate in months. The things have gotten reliable enough that it's almost more plausible to me that someone's making stuff up, and using that as input for the LLM, and then blaming the LLM for the fabrications.

Over-confident? Still a problem. Sycophant? Still a problem. Able to be confused with garbage input? Still a problem. Making stuff up out of nowhere? I haven't seen that in quite a while... Frankly, I'd want to see the chat log before believing it.

~D

What a time to be alive!

When people tell me that AI is just a "plagiarism machine", I point to stuff like this and the AlphaFold database... Advances in science, mathematics, and research that have been out of our grasp for decades... Solved in novel ways.

~D

They took care of content distribution. They took care of international sales / taxes / etc. They took care of currency conversions. They took care of folks who wanted returns (not many, but still).

They made my game accessible world-wide with almost zero effort on my part.

Did I give them a 30% commission to sell my game? Yes. Would I like that percentage to be lower? Sure. Would I call it excessive? Hell no. Especially since retail stores get a similar cut, and I didn't have to worry about manufacturing, distribution, stock, returns, etc.

~D

I disagree. Some sites, like allrecipes, are better than the Google summary. I'd much rather have thorough content that is convenient to access, then impartial content, but I would prefer impartial content over thorough content that is a pain in the ass to access

I gotta agree with this take...

As an old-school web developer, one of the things that the website must, absolutely, do, is be more convenient than the alternatives.

The AI summary is quick, available at the top level, and is of inferior quality. Scrolling by it, if it's not what you want, is trivial. We've scrolled by sponsored links at the top of search results for decades.

The website article contains a lot more friction: Getting past the ads and the narrative schlock to get to the actual recipe. It's far more convenient to use a "good enough" recipe, instead of wade through the garbage on these sites... To get the parts of that site that wouldn't be eligible for copyright protection if they hadn't made things crappier (which is what all that prose is for).

It's enshittification for profit, and I'm not surprised the intended audience is looking for other ways to get similar information.

~D

This. I'm a working professional, and I also teach. Truth is, the industry I work in changes significantly from year to year... I'm constantly updating content in my lectures, and these tools can be fantastic support, but they're not good instructors.

"nearly a quarter were using AI tools in their teaching"
Hell yes I do, it'd be insane for me to not use them. AI is great for "Did I forget anything?" "Is there a contrarian approach that I'm ignoring?" or "Make this awkward paragraph read more clearly"

I also expect students to use AI tools to help them dig into side topics that I can't get into during class time. I expect them to use it to improve the quality of their projects. (I'm in a position where I teach other professionals, who are learning the topics because they -want- to know the content - we don't have grades)

I absolutely use these tools to make sure that my course content is well-rounded and professionally done. But having an AI generate the slides carte blanche and then teach it? No thanks. There is value in having a human teacher that actually understands the weight and value of some topics over others.

Because the professor could be faking it by using ChatGPT, and be providing incomplete/wrong information. The fact that the -student- noticed problems in the course material means that she was not getting the expertise and precise information exchange expected of someone with the title of "Professor"

I think you're right, and I also think you're hitting on the key takeaway:

ChatGPT let someone solve a problem quickly, but the person didn't learn the skill.

So, when faced with a task, we should ask ourselves: Do we want to solve this quickly, or do we want to learn the skill?

It used to be we always had to learn the skill to solve the problem. It's interesting that now we have a choice.

Since I, as an individual, have no interest in being a FORTRAN programmer, I'm OK with not learning the skill. Sometimes you just need to solve a thing to move on with your life, so you can invest the time and energy on the things you care about.

~D

The retroactive portion was the most troubling bit, because it was a violation of their own TOS. With these new adjustments, the new fees only apply starting with the next LTS release of the engine, so it's no longer retroactive. That makes the math far easier, since it's only for new projects.

~D

The biggest problem is, by looking at the drive from the outside, there is no way to distinguish which devices have critically sensitive data, and which have been properly wiped. That means there is always room for human error as drives get moved around, or get mixed up, and non-wiped devices could get put back into use (or even worse, sold). That is: It's a human factors problem, not a technical one.

Shredding doesn't have that ambiguity.

~D

3D modeling, design review, training, scientific visualization and historic site preservation are just a few non-entertainment uses of VR which have proven to be very valuable.

But even as a VR professional, I think working all day in-headset is stupid.

~D

As a matter of practice, I don't publicly associate personal accounts and my job. You can e-mail my slashdot username at yahoo, and I'll answer your question personally.

~D

I work at a research organization. Many of the hot research topics aren't even fully formed yet. When I interview, I don't ask folks for what skills they have, I ask where passions lie. Folks who are passionate will clearly demonstrate it, and they will grow into the roles they're interested in. Folks who don't have passion, don't make it. I can train just about anyone, except people uninterested in being trained.

~D

So I also have made indie games for the public, but didn't put nearly as much of myself into it, and I have nothing but respect for you. Even with a small game, it's really tough... What I realized through the process is that I liked ~making~ games which is a very different activity from ~selling~ games, which I turn out to hate. I haven't found a fantastic way to navigate beyond that for small indie games, as many publishers are looking for full-time devs.

~D

There's also additional market pressures which Microsoft may be able to leverage that Google wasn't able to. Right now new gaming equipment is still hard to come by, and costs a premium. With older xBox systems (an audience Google couldn't tap) being able to play newer games, it provides MS with a specific audience of folks who may be looking for stop-gap measures to play the latest games until they can find/afford the latest hardware.